Keith Lofstrom wrote: > I am also curious about XFS and JFS and other alternatives. I get the > impression that all of these next-generation journalling file systems > are somewhat experimental. John Thompson wrote: > XFS is hardly experimental; it has been the primary filesystem for SGI's > IRIX operating system for over a decade. > > JFS is somewhat newer, having been developed for IBM's OS/2 operating > system within the last decade and more recently ported to linux. But > IBM has many, many years of experience designing filesystems and JFS has > certainly benefited from that. Good point. I should have said "... the integration of these next-generation journalling file systems with recent Linux kernels is somewhat experimental". I'm sure the underlying architecture of both of these file systems is quite sound and well tested, but as one can see from perusing the SGI site, the Linux kernel and IRIX are different enough to change some of the assumptions on which XFS is based. I assume the same is true for OS/2 and Linux . While I am sure that both the SGI and IBM teams do world-class work, I wonder how much testing these file systems get relative to ext2/3 , especially with fedora-tweaked kernels. Hence my question. The same question holds for reiserfs, of course. In what circumstances do these file systems break? Keith -- Keith Lofstrom keithl@xxxxxxxxxx Voice (503)-520-1993 KLIC --- Keith Lofstrom Integrated Circuits --- "Your Ideas in Silicon" Design Contracting in Bipolar and CMOS - Analog, Digital, and Scan ICs
